This tax case involves a tax avoidance strategy at its heart, one of the first marketed tax strategies to come before the First Tier Tribunal involving Gilt strips, Andrew Berry v Revenue & Customs. There were very many versions of this planning at the time and it is likely that hundreds of people took part and have been waiting in anticipation for the result. This particular version was sold by Abacus Wealth Planning (“Abacus”) and the bank involved was SG Hambros.
So what are Gilt strips?
Well, a Gilt strip is a UK Treasury Government stock. The word “strips” is an acronym that has come to mean “Separate Trading of Registered Interest and Principal Securities”. A three year Gilt can, for example, be broken down into 7 Gilt strips. This is because a three year Gilt will entitle its holder to six interest payments (one every six months) and the repayment of the principal at the end of the three years. Each one of these seven cashflows can be “stripped” from the Gilt to form Gilt Strips which can each be separately bought and sold.
In order to understand the decision of the Tribunal it is important to understand the planning at least in outline and in particular the timelines and the cashflows.
In brief, the following transactions took place, all by way of a power of attorney signed by Mr Berry in favour of four individuals that worked for Abacus:
- On Day 1, Mr Berry entered into a Forward Contract to buy approximately £6.5m of Gilt strips. He had until 5pm on Day 4 to deliver a purchase notice to actually acquire the strips;
- Also on Day 1, Mr Berry sold an option to a third party to acquire the strips. The third party paid £390,000 for the option which required it to buys the strips for £6.1m if exercised;
- On Day 4 at around 4.45pm, the purchase notice was signed and presented to the bank. The strips were paid for by way of a loan of £6.1m from the bank plus the £390,000 option price plus the fees Mr Berry paid for the planning. The strips were then transferred to Mr Berry;
- On Day 4 at around 5.15pm, the option was exercised. Mr Berry was paid £6.1m which was used to repay his bank loan and the strips were then transferred to the third party.
Following the Scottish Provident Institution case in 2003, Abacus were advised by Tax Counsel that there had to be a real chance that the value of the Gilt strips would fall below the option strike price thereby leaving the client with the Gilt strips because the third party would not exercise the option. A statistical analysis showed that the chances of this happening between Days 1 and 4 were 7%. This was accepted by the Tribunal. The problem was that the strategy was then designed so as to negate even that risk.
Firstly, although Mr Berry entered the Forward Contract on Day 1, he did not have to actually serve the purchase notice until Day 4. So his risk was in actual fact only for the 30 minutes between 4.45 pm and 5.15 pm on Day 4. Secondly, if the strips value fell below the option strike price a get out had been built in for Mr Berry. He could decide not to serve the purchase notice and in return for this default he could instead pay an amount equal to the option price of £390,000 plus his fees.
Therefore if the “7% risk” ever actually materialised, he would only be down by his fees and Abacus even agreed to re-do the planning, if still available, for free. The Tribunal therefore decided that, “...any risk...to any of the participators in the Gilt Strip Planning was a façade. It had no reality. It was built into the Gilt Strip Planning scheme but the parties proceeded on the basis that it should be disregarded.”
In the Relevant Discounted Securities case, Astall v HMRC 2009, a similar point was made when the case was heard (and lost by the taxpayer) in the Court of Appeal. This case involved the creation of a loss for income tax purposes using Relevant Discounted Securities (RDS) and again the scheme designers had included risk elements. One of these was designed to ensure that the debt instrument used in the structure was an RDS, an element vital to the planning. In simple terms an instrument is an RDS if when it is redeemed, a profit in excess of a certain percentage of the original issue price might be made. This is known as a “deep gain”. It isn’t necessary for this “deep gain” to actually be made on redemption but it must be a possibility.
The instrument used in the Astall case therefore provided that a “deep gain” would arise if a “market change condition” was not met and the client decided to wait until the maturity date in 15 years time. The chance of the “market change condition” not being met was said to be 15% and that was not disputed. However, as in the Berry case the client had a get out clause in that he could, if the risk materialised, simply redeem the instrument within the initial two months and did not need to wait until maturity. As a result the Court of Appeal judges were simply not convinced that there was any real chance of a “deep gain” arising and hence decided that the instrument was not an RDS. Arden LJ stated that,
“…there should be a real possibility of a deep gain if losses incurred on a RDS were to be offsettable for income tax purposes. In this case, terms of issue which on the face were essential for the securities to qualify as RDS had to the extent described above no practical reality and must therefore be disregarded.”
These two cases highlight that it is not sufficient simply to include risk elements. These risk elements must have a reasonable chance of actually happening and must not be rendered impotent and hence ignorable by the inclusion of safety nets for clients. This could make the design and marketability of some avoidance strategies much more difficult as many clients will not be prepared to take the commercial risks needed to make them effective.
Finally, the Tribunal had two further points to make which may well be welcomed by people that have taken part in planning that has yet to be litigated or intend to do so in the future particularly if these do not include any artificial risk factors as seen in Berry and Astall.
Firstly despite highlighting a number of flaws in both the implementation and design of the planning, the Tribunal rejected the HMRC argument that the any part or the whole of the arrangements were a sham. They were satisfied that despite the fact that there was “...no tax reality...” there was “...sufficient legal reality to displace a conclusion of sham.”
And secondly, the Tribunal reiterated the point made in Barclays Mercantile Business Finance Ltd v Mawson, that a strategy may unashamedly have a tax avoidance motive and may even have steps inserted simply to advance that motive but absent specific anti-avoidance legislation, this is quite irrelevant when deciding whether legislation does or does not apply.